DOI: /peds The online version of this article, along with updated information and services, is located on the World Wide Web at:

Transcription

1 Direct Cost Analyses of Palivizumab Treatment in a Cohort of At-Risk Children: Evidence From the North Carolina Medicaid Program Steven Wegner, Julie Jacobson Vann, Gordon Liu, Patricia Byrns, Clement Cypra, William Campbell and Alan Stiles Pediatrics 2004;114;1612 DOI: /peds The online version of this article, along with updated information and services, is located on the World Wide Web at: PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly publication, it has been published continuously since PEDIATRICS is owned, published, and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois, Copyright 2004 by the American Academy of Pediatrics. All rights reserved. Print ISSN: Online ISSN:

2 Direct Cost Analyses of Palivizumab Treatment in a Cohort of At-Risk Children: Evidence From the North Carolina Medicaid Program Steven Wegner, MD, JD* ; Julie Jacobson Vann, PhD*; Gordon Liu, PhD ; Patricia Byrns, MD ; Clement Cypra, MHA ; William Campbell, PhD ; and Alan Stiles, MD ABSTRACT. Objective. Use of palivizumab prophylactic therapy reduces the occurrence of hospitalizations for serious respiratory syncytial virus (RSV) lower respiratory tract infections in at-risk infants. The direct costbenefit of palivizumab prophylaxis for infants who are born at 32 to 35 weeks estimated gestational age (EGA) during their first year of life has not been systematically examined. The objective of this study was to compare the direct costs of palivizumab prophylaxis and RSV treatment in infants who were born at 32 to 35 weeks EGA and received and did not receive palivizumab. Methods. A cohort study was performed of infants who were younger than 1 year and were enrolled in an enhanced primary care case management model within the North Carolina Medicaid Program. Comparisons were made between infants who received (Synagis prophylaxis group) and did not receive palivizumab (nonprophylaxis group) during the study period. Cost was examined using the sum of Medicaid paid services for prophylaxis with palivizumab and treatment for RSV infections that occurred between October 1, 2002, and May 31, The Anderson framework was used to specify the regression cost models to compare the participants who received (Synagis prophylaxis) and did not receive (nonprophylaxis group) palivizumab. The primary outcomes were actual 7-month seasonal costs and standardized seasonal costs adjusting for the varied months of infant participation. Results. The study sample included 185 Synagis prophylaxis and 182 nonprophylaxis participants who met the inclusion criteria. The average per-person total cost of RSV care and prophylaxis was $5117 for the Synagis prophylaxis group and $371 for the nonprophylaxis group. Five hospitalizations occurred in the prophylaxis group, and 12 occurred in the nonprophylaxis group (odds ratio: 0.27). No deaths occurred in either group. Conclusions. Palivizumab administered to infants who were born at 32 to 35 weeks EGA did not provide direct cost savings related to hospitalization or ambulatory care in a Medicaid population. The primary difference in cost between the groups was attributable to the palivizumab prophylaxis. Pediatrics 2004;114: ; From *AccessCare, Morrisville, North Carolina; and Department of Pediatrics, School of Medicine, Pharmaceutical Policy and Evaluative Sciences, School of Pharmacy, Department of Medicine, School of Medicine, and Department of Health Policy and Administration, School of Public Health, University of North Carolina, Chapel Hill, North Carolina. Accepted for publication Jun 29, doi: /peds No conflict of interest declared. Reprint requests to (S.W.) AccessCare, 3500 Gateway Centre Blvd, Suite 130, Morrisville, NC sew@ncaccesscare.org PEDIATRICS (ISSN ). Copyright 2004 by the American Academy of Pediatrics. respiratory syncytial virus, bronchiolitis, prophylaxis, palivizumab, cost analysis, Synagis. ABBREVIATIONS. RSV, respiratory syncytial virus; EGA, estimated gestational age; CLD, chronic lung disease; AAP, American Academy of Pediatrics; PA, prior authorization. Respiratory syncytial virus (RSV) is the leading cause of serious lower respiratory tract infections in children. 1 7 Serious RSV disease, often measured by the need for hospitalization, is predominantly manifested as bronchiolitis in infants yet may occur as pneumonia. 5 In the United States, hospitalizations are attributed to RSV per year. 6,8,9 The efficacy of palivizumab, a humanized monoclonal antibody, was investigated during the RSV season in the Impact-RSV randomized, double-blind, placebo-controlled trial. 6 Participants who received palivizumab experienced a 55% reduction in RSV hospitalizations compared with nontreated control subjects for all risk groups in the study sample. In this same randomized trial, an 80% reduction in RSV hospitalizations was observed in the 32- to 35-week estimated gestational age (EGA) subgroup. In contrast, an observational study that examined children who were 10 months or younger at the start of the season and had a diagnosis of prematurity or chronic lung disease (CLD) and were enrolled in the Kansas Medicaid Program reported that the impact of palivizumab on hospitalization reduction was smaller (50%). 10 In 1998, the Food and Drug Administration approved the drug palivizumab for preventing serious lower respiratory tract disease caused by RSV in pediatric patients who are at high risk for RSV disease. 4,11 That same year, the American Academy of Pediatrics (AAP) recommended consideration of palivizumab use for infants (1) who are younger than 2 years with CLD requiring medical therapy for this condition within 6 months of the RSV season; (2) who were born at 28 weeks EGA or earlier up to 12 months of age; (3) who were born at 29 to 32 weeks EGA for up to 6 months of age; and (4) who were born between 32 and 35 weeks EGA only with additional risk factors defined as underlying conditions that predispose to respiratory complications (eg, neurologic disease in very low birth weight infants), number of young siblings, child care center attendance, exposure to tobacco smoke in the home, anticipated cardiac surgery, and distance to and 1612 PEDIATRICS Vol. 114 No. 6 December 2004

3 availability of hospital care for severe respiratory disease. 11 The AAP published revised recommendations for RSV prophylaxis with palivizumab in The 2003 recommendations for infants in the 32- to 35-week EGA risk group reserve palivizumab prophylaxis for those with 2 of the following risk factors: child care attendance, school-aged siblings, exposure to environmental air pollutants, congenital abnormalities of the airways, and severe neuromuscular disease. 12 The North Carolina Medicaid program has used a series of measures to encourage compliance with the 1998 AAP published recommendations, including implementation of a prior authorization (PA) program for palivizumab use. During the RSV season, the North Carolina PA criteria closely matched the 1998 AAP published recommendations. Participants who met the following guidelines were automatically approved for palivizumab administration: (1) infants who were 24 months with CLD that has necessitated treatment in the last 6 months, (2) infants who were born at 28 to 32 weeks EGA without CLD and were younger than 6 months at the start of the RSV season, (3) infants who were born at 28 weeks EGA without CLD and were younger than 12 months at the start of the RSV season, and (4) infants who were born at 32 to 35 weeks EGA without CLD and were younger than 6 months at the start of the RSV season with 2 illnesses or risk factors that predispose the infant to respiratory complications. Any requests that were sent back to the primary care physician requested missing or additional information. All requests that were resubmitted were ultimately approved. The cost benefit of palivizumab use in infants who were born at 32 to 35 weeks EGA and are younger than 1 year has been studied with mixed results. 8,10,13 16 Joffe et al 13 reported cost estimates of $ to $ per hospitalization averted for RSV infection in infants who were born at 32 to 35 weeks EGA, taking into consideration the duration of perinatal oxygen use and month of neonatal intensive care unit discharge. Shireman et al 15 reported that in all patients, costs of prophylaxis with palivizumab were 6.67 times higher than RSV treatment costs in the Kansas Medicaid program. A number of published cost analyses used hypothetical cohorts with published costs and hospitalization rates from efficacy trials, and several applied hospitalization costs or charges that were not specific to the 32- to 35-week EGA subgroup Concern exists over the high costs associated with palivizumab prophylaxis and the absence of definitive peer-reviewed economic evaluations of palivizumab use in the 32- to 35-week EGA subgroup. 11,13 The cost of palivizumab prophylaxis (drug only) to the North Carolina state Medicaid program was $12.3 million in The aim of this investigation was to determine whether the costs of palivizumab use in the 32- to 35-week EGA subgroup outweigh the direct health care costs associated with serious RSV infection in a cohort of North Carolina Medicaid recipients who qualify for palivizumab treatment under current guidelines. Costs and impact on participants health after the study period were not examined. METHODS Patient Sample A cohort study was designed to compare health care costs and utilization associated with inpatient and ambulatory care treatment of and prophylaxis against RSV infections and RSV-attributed bronchiolitis between 32- and 35-week EGA participants who received (Synagis prophylaxis) and did not receive (nonprophylaxis) palivizumab (Synagis) during the RSV season. 17 The observational study period was October 1, 2002, through May 31, Palivizumab is typically administered monthly beginning just before the onset of the RSV season (October to December), with the last dose given at the end of the RSV season (March to May). 7,11,12 The timing of the RSV season varies by region within the United States. 11 During the study period, the North Carolina Medicaid program approved up to 6 doses of palivizumab per patient between October and March, with some doses having been approved for administration during April The study design was reviewed and approved by the University of North Carolina School of Medicine Human Subjects Institutional Review Board and the East Carolina University and Medical Center Institutional Review Board. The Anderson model for access to health care services was selected as the theoretical framework to guide identification of covariates that have the potential to influence patients access to and utilization of health care services. This model categorizes characteristics of the population at risk into 3 groups: predisposing, enabling, and need (risk) variables. 18,19 The predisposing variables are those that describe the propensity of individuals to use services, such as gender, race, religion, and values regarding health. 18 The variables race, gender, and age of participant s mother at the infant s date of birth were selected as predisposing variables to be used in our models. The enabling variables refer to resources that the patients and families have to use services, such as income, insurance, and geographic region. 18 The enabling variables incorporated into our models include the highest educational level attained by the infant participant s mother and the number of miles the infant lived from a hospital that cares for infants or children or has an emergency department. 11,12 Because all study participants were covered by Medicaid during the study period, income and insurance status were not selected for statistical modeling. The need variables refer to one s illness level or perceived need for care as evaluated by the individual or the health care delivery system. 18 The need variables were selected for this study on the basis of the North Carolina PA guidelines for palivizumab use in and include whether an infant participant (1) had a sibling in school, (2) attended day care during the study period, (3) is exposed to cigarette smoke in the home, or (4) was part of a multiple birth. 11,12 Risk variables identified from the North Carolina PA program were supplemented with the number of people living in the infant participant s home during the study period (as a proxy for crowding), the infant s estimated gestational age in weeks, and the infant participant s birth weight in kilograms. 11 Patients for the study were selected from all 28 widely dispersed pediatric practices that serve North Carolina Medicaid patients within the AccessCare/Community Care Program, an enhanced primary care case management program. Study participant recruitment began in late January 2003, independent of the physicians decisions to recommend palivizumab and of the North Carolina PA program. Potential study participants, those having received or not received palivizumab, in any of the North Carolina PA risk categories were identified by local care managers at participating practices through review of palivizumab PA forms, Medicaid claims, birth lists and medical records, and meetings with Synagis nurses. Potential participants were screened for eligibility using date of birth, EGA, Medicaid and Carolina Access enrollment, and comorbidities. Patients were eligible for this economic study when they were born between March 1, 2002, and February 28, 2003, at 32 to 35 weeks EGA and enrolled in the AccessCare Medicaid program during the study period. Patients were excluded from this 32- to 35-week EGA subgroup when they had a diagnosis of CLD and/or hemodynamically unstable congenital heart disease to study a relatively homogeneous risk ARTICLES 1613

4 group. CLD was defined as patients with CLD who have required medical therapy (supplemental oxygen, bronchodilator, diuretic, or corticosteroid therapy) for CLD within 6 months before the anticipated start of the RSV season. 7 A physician and nurse team reviewed patient diagnoses and treatments listed in medical records and Medicaid claims data to identify patient exclusions, as defined in the Red Book, Data Collection Study data for Synagis prophylaxis and nonprophylaxis participants were collected from parental survey, medical record abstraction, and North Carolina Medicaid claims and demographic files. Local care managers provided names of eligible children and contact information to research personnel. Parents or guardians were contacted by research staff at the end of the season by telephone to perform a 21-question survey to identify risk factors for RSV hospitalization, respiratory-related hospitalizations, and demographic data. When families were not reached by telephone, data were collected by home or clinic visit or chart review. Survey data were entered into a SQL server database. Medicaid health care claims were obtained in December 2003 for health care services that were provided during the study period. Inpatient claims were screened using diagnoses: acute bronchiolitis due to RSV, acute bronchiolitis due to other infectious organism, acute bronchitis, RSV pneumonia, pneumonia organism not otherwise specified, and viral pneumonia. For inpatient stays identified by claims and/or survey, care managers abstracted medical records to determine RSV test results (rapid antigen tests). Hospitalizations and related costs were included in the analyses when RSV tests were positive or when RSV tests were not performed but RSV was listed in the final diagnoses. When RSV tests were negative, hospitalizations were excluded. Ambulatory care claims, including outpatient, emergency department, and office visits, were included for all RSV-related and/or bronchiolitis diagnoses or wheezing because RSV testing is not routinely performed by providers in this study for ambulatory care services. Ambulatory care claims were not adjusted for lack of RSV testing because previous cost studies of palivizumab have found cost modeling to be insensitive to changes in the number of emergency department and physician office visits. 14 Palivizumab injection dates and doses were abstracted from medical records, including injections that were given during hospitalizations. Palivizumab was not provided to study participants during home visits. Palivizumab injection data (medical records) were compared with Medicaid palivizumab drug claims. Medical records data were selected as the source for dates of administration and doses because palivizumab is provided to the physician practice and billed by a pharmacy provider that is a wholesale distributor within the North Carolina Medicaid program. In this program, the palivizumab drug claim service dates frequently precede the injection dates in the medical records. The costs of injections were assigned to injections on the basis of doses and North Carolina Medicaid reimbursement rates. When only the injection date was abstracted from the medical record, the cost was assigned from the claims database. Palivizumab costs were adjusted for (1) the manufacturer s discount to the state Medicaid programs according to the federal rebate program, (2) drug handling fees, and (3) nurse visits (in lieu of physician office visits billed for some injections) for conservative cost estimates. Participants were assigned to the nonprophylaxis study group in the absence of documented evidence of palivizumab injections and to the Synagis prophylaxis group when there was documented evidence of receipt of at least 1 palivizumab injection during the study period. RSV-attributed hospitalizations were reviewed for temporal relationships to palivizumab injections to identify prophylaxis status at the time of admission. The 4 participants who were identified as having RSV-attributed hospitalizations occurring before receipt of palivizumab were assigned to the nonprophylaxis study group as this was their prophylaxis status as of the time of admission. All eligible health care costs and utilization for these 4 participants were assigned to the nonprophylaxis study group for conservative cost estimates. This included the costs of 1 participant who was hospitalized with an RSV infection both before and after palivizumab prophylaxis, whose costs were not allocated between the 2 study groups because this report does not include person-time analytic methods. Statistical Methods In these analyses, 2 primary outcome variables were defined: hospitalizations for RSV-related conditions and the total RSVrelated health care costs per individual for the 7-month study period. Taking the payer perspective, in this case North Carolina Medicaid, the total RSV-related cost was defined to include expenditures on all RSV-related health care services including hospitalization, emergency department, ambulatory care (outpatient and office visit), and palivizumab prophylaxis costs. Total seasonal costs were measured in both actual value and standardized value. The latter was used to account for the fact that some infants were born after the beginning of the study period and therefore did not complete the full 7-month study period. The total seasonal cost variable was standardized by dividing the total individual costs by the actual number of months that a child was in the study and multiplying by 7 months. Because the rapid antigen tests used to detect RSV have sensitivities that are 80% to 90% effective, 20 it is likely that all RSV-attributed hospitalizations were not captured using the inclusion criteria. Therefore, sensitivity analyses of total costs were performed by including hospitalization costs for non- RSV bronchiolitis for comparison. Power calculations estimated that a sample size of 16 in each study group was needed to detect an effect size of 1.92 or greater using projected average monthly costs (outcome) with 90% power (.05 level of significance, 1-sided). 21 For detecting differences in hospitalization proportions with a relative risk of 0.2, at least 114 patients per study group were required. 22 Using the Anderson framework, 18,19 a descriptive analysis was conducted to compare predisposing, enabling, and need factors between the Synagis prophylaxis and nonprophylaxis study groups, with differences tested using the 2 test for dichotomous and categorical variables and the t test for continuous variables. 23,24 The pooled t test was selected for equal variances, and Satterthwaite t test was chosen for unequal variances, based on the F test. The descriptive statistics provide information on possible confounding of palivizumab s protective effect as a result of potential sample selection bias. Unadjusted per-participant average seasonal costs and utilization were compared between the study groups for RSV-attributed hospitalizations, ambulatory care visits, emergency department visits, and palivizumab injections. In addition, to assess the similarity of health status between the 2 study groups, the overall health care costs for the duration of the study period, net of the costs of RSV treatment and palivizumab prophylaxis were compared. After the descriptive analysis, multivariate regression analyses were conducted to assess the protective effect of palivizumab on the risk of hospitalization and costs while controlling for possible confounding factors, such as previously identified risk factors. Specifically, 3 multivariable regression models were estimated. Model 1 is a logistic regression model that estimated the betweengroup difference in the likelihood of RSV hospitalization during the study period. Model 2 is a linear regression of individual seasonal total costs as a function of palivizumab intervention and other covariates, using both actual and standardized costs. Model 3 is a log-transformed regression of the seasonal costs in an attempt to minimize the highly skewed distribution of the cost data. In each regression model, the potential influence of risk factors and other covariates on costs and hospitalizations, individually, as index variables, and as dichotomized index variables, was assessed. Index risk factor variables were created by counting the number of positive risk factors identified for each study participant, first using PA criteria and second using PA criteria. Dichotomous risk factor variables then were created by identifying participants with 2 positive risk factors versus 2. RESULTS Of the 714 patients who were identified as being potentially eligible to receive palivizumab on the basis of the North Carolina PA guidelines, 374 were born after February 28, 2002, at 32 to 35 weeks EGA. Of these, 6 were excluded for hemodynamically unstable congenital heart disease and 1 was excluded for CLD, leaving 367 who met eligibility criteria for this cost analysis (Table 1). Surveys 1614 COST ANALYSIS OF PALIVIZUMAB PROPHYLAXIS

5 TABLE 1. Study Participant Enrollment, by Study Group Description Prophylaxis Nonprophylaxis Total Potential participants in all risk groups 714 Infants 32 wk EGA (excluded from this report) 340 Potential participants wk EGA wk EGA; excluded for CHD wk EGA; excluded for CLD Included 32- to 35-wk EGA participants CHD indicates congenital heart disease. were fully or partially completed for all 367 study participants, with 86.6% (318) completed by telephone or in person and 13.4% (49) by medical record abstraction. Risk factor data were complete for 361 (98.4%) study participants. Study participant attributes and risk factors were statistically similar between study groups, on average, for patient gender, mother s age, distance to a health care facility that cares for children, having siblings in school, smoking exposure in the infant s home, and the number of positive risk factors per participant on the basis of the 2002 PA criteria (Table 2). However, the Synagis prophylaxis group had a higher concentration of white infants (53.5% vs 34.1%; P.0002), had a lower percentage of black (36.8% vs 51.6%; P.004) and Hispanic (1.1% vs 7.1%; P.003) infants, had a higher percentage of multiple births (33% vs 15.9%; P.0001), and were of slightly lower gestational age (33.4 vs 34.0 weeks; P.0001) and weight (1.99 vs 2.20 kg; P.0001) at birth compared with nonprophylaxis participants. Being part of a multiple birth was associated with the dependent and prophylaxis variables (P.0001) in this study sample. The average per-participant costs of all non RSV-related treatment and prophylaxis during the study period were $4820 for the Synagis prophylaxis group and $5184 for the nonprophylaxis group. Risk of RSV-Related Hospitalization During the RSV season, a total of 17 hospitalizations with a positive RSV test and/or diagnosis were identified in the study sample (Table 3). Twelve of the RSV hospitalizations were observed in the nonprophylaxis participants (hospital proportion: 0.066), and 5 were observed in the Synagis TABLE 2. Patient and Family Attributes and Risk Factors of Infants Who Were Born at 32 to 35 Weeks EGA, by Study Group, Using the Anderson Framework Variables Variable Type Prophylaxis Mean Value (n 185) Nonprophylaxis Mean Value (n 182) Predisposing Infant race White Dichotomous 53.51% 34.07%.0002 Black Dichotomous 36.76% 51.65%.0041 Hispanic Dichotomous 1.08% 7.14%.0034 Indian Dichotomous 3.78% 3.85%.9751 Asian and other Dichotomous 4.86% 3.30%.4481 Male gender Dichotomous 49.19% 46.15%.5605 Mother s age Continuous y y.5057 Enabling Mother s education No formal school Dichotomous 0.56% 0.61%.9505 Elementary school Dichotomous 2.23% 4.88%.1832 Middle school Dichotomous 3.91% 2.44%.4397 High school Dichotomous 70.39% 59.76%.0387 College or graduate school Dichotomous 22.91% 32.32%.0509 Miles from health care facility Continuous 12.0 miles 13.7 miles.1401 Dichotomous 13.66% 18.68%.1927 Need (Risk) Siblings in school Dichotomous 44.57% 36.46%.1150 Attend child care Dichotomous 16.94% 23.89%.1004 Smoke exposure Dichotomous 19.02% 21.98%.4836 Multiple birth Dichotomous 32.97% 15.93%.0001 No. of people in home Continuous No. of positive risk factors, Continuous EGA Continuous wk EGA Dichotomous 28.11% 13.19% wk EGA Dichotomous 24.32% 17.58% wk EGA Dichotomous 28.65% 30.22% wk EGA Dichotomous 18.92% 39.01%.0001 Birth weight Continuous 1.99 kg 2.20 kg.0001 VLBW Dichotomous 12.97% 7.18%.0662 LBW Dichotomous 75.14% 65.75%.0492 NBW Dichotomous 11.89% 27.07%.0002 VLBW indicates very low birth weight ( 1500 g); LBW, low birth weight ( g); NBW, normal birth weight ( 1500 g). ARTICLES 1615

6 TABLE 3. Utilization of RSV-Related Health Care Services for Eligible 32- to 35-Week EGA Infants, by Study Group Comparison Prophylaxis Group (n 185) Nonprophylaxis Group (n 182) Total no. of hospitalizations Average no. of ambulatory care visits per participant Average no. of emergency department visits per participant Average no. of palivizumab injections per participant prophylaxis group (hospital proportion: 0.027). Nonprophylaxis hospitalizations included 5 that occurred in 4 patients with RSV hospitalizations preceding RSV prophylaxis. One of these 4 participants was hospitalized before and after RSV prophylaxis. Synagis prophylaxis participants were hospitalized a total of 22 days compared with 45 days for nonprophylaxis. Prophylaxis participant hospitalizations included 2 days in the intensive care unit compared with 13 days for nonprophylaxis participants; 9 of 13 days were for 1 hospitalization. When controlling for all observed covariates as possible confounding factors, the multivariate logistic regression shows that prophylaxis and nonprophylaxis participants were not different for incidence of hospitalization for RSVrelated conditions (odds ratio: 0.27; P 0.058). In the hospitalization regression models, none of the risk factors altered the relationship between RSV hospitalization and palivizumab prophylaxis. No deaths were reported in either the prophylaxis or the nonprophylaxis study group. Costs of Care and Total Cost Analysis The aggregate seasonal cost of RSV inpatient hospitalizations for both study groups combined was $57 028, including 1 outlier that cost $21 955, at an average hospitalization cost of $3355. At the individual participant level, the average seasonal cost of RSV hospitalizations in the nonprophylaxis group ($250; standardized: $361) exceeded costs in the prophylaxis group ($64; standardized: $76); this difference was not statistically significant (P.16; Table 4). For ambulatory care only, including outpatient and office visits, the average seasonal costs for bronchiolitis and RSV-related care were $54 (standardized: $60) for nonprophylaxis and $46 (standardized: $50) for Synagis prophylaxis participants (P.62). Study participants received 770 palivizumab injections in aggregate during the study period, at an average cost of $1214 per injection. Adding all cost components together, the mean per-person seasonal total costs of RSV-related care and prophylaxis was $5117 (standardized: $5434) for the Synagis prophylaxis participants and $371 (standardized: $505) for nonprophylaxis participants (P.0001). When the total cost analyses were repeated by adding non-rsv bronchiolitis hospitalizations, the average per-person seasonal total costs of RSV-related care and prophylaxis were $5161 for the Synagis prophylaxis group and $524 for the nonprophylaxis group. Holding all other observed covariates constant, the multivariate cost model indicates that the Synagis prophylaxis participants had per-person costs that exceeded those of the nonprophylaxis participants by $4634 (P.0001) for the 7-month season (Table 5, full model). Alternatively, when the regression analysis was repeated with a log transformation of the dependent variable, the Synagis prophylaxis participants were found to have an average total cost that was 7 times greater than the cost incurred by nonprophylaxis participants (P.0001). In each regression cost model (Table 5), only the multiple birth variable was found to be significantly associated with increased costs. The other risk factors examined, including miles from a health care facility, having siblings in school, day care attendance, and ex- TABLE 4. Actual and Standardized Per-Person Seasonal Costs of RSV-Related Health Care Services, by Study Group, for all 32- to 35-Week EGA Study Participants Comparison Prophylaxis Group, $ (n 185) Nonprophylaxis Group, $ (n 182) Average seasonal cost of inpatient services Actual Standardized* Average seasonal cost of ambulatory care services Actual Standardized* Average seasonal cost of emergency department Actual visits Standardized* Average seasonal cost of palivizumab injections Actual Average seasonal costs of treatment and prophylaxis Standardized* Actual Standardized* All costs are per-participant seasonal costs. * Standardized costs actual per-person cost, divided by number of months in study, multiplied by 7 months. Four patients who received palivizumab injections after RSV hospitalization were classified as nonprophylaxis participants COST ANALYSIS OF PALIVIZUMAB PROPHYLAXIS

7 TABLE 5. Model 2: Multivariate Regression Analysis: Dependent Variable Total Seasonal Costs of RSV Treatment and Prophylaxis Variable F Model Coefficient F Model Sumrisk1 Model* Coefficient Sumrisk1 Model* Sumrisk2 Model Coefficient Sumrisk2 Model Palivizumab Black Hispanic Indian Asian/other Male Mother s age Mother college Mother high school Mother middle school Mother elementary school Miles from health care facility Siblings in school Day care attendance Smoke exposure Multiple birth No. in home EGA Birth weight Sumrisk Sumrisk df R Total seasonal costs include expenditures that occurred in October 2002 through May 2003 for inpatient hospital, outpatient, emergency department, ambulatory care, and palivizumab intervention. * Includes the Sumrisk1 variable, an index created from the 5 risk factors identified in the PA program. Includes the Sumrisk2 variable, an index created from 3 risk factor variables selected in the PA program. posure to cigarette smoke in the home, were not found to alter the relationship between costs and palivizumab prophylaxis in this study sample. The risk factor index variables, Sumrisk1 (number of positive risk factors using criteria) and Sumrisk2 (number of positive risk factors using criteria), were not statistically significant in any of the hospitalization or cost models (Table 5). In addition, the dichotomous risk factor variables ( 2 positive risk factors) were not statistically significant in any of the regression models. DISCUSSION Previous cost studies of palivizumab have been criticized for omitting social, family, and potential long-term costs; not thoroughly quantifying costs of palivizumab administration and ambulatory care services; and insufficient statistical power. 20 Our investigation addressed several but not all of these limitations; however, our findings were robust in all sensitivity analyses performed. One limitation of our study was our focus on direct health care costs of RSV prophylaxis and treatment to the North Carolina Medicaid program during the study period only. This study did not consider other societal costs or possible long-term health effects of RSV bronchiolitis. A study conducted in Sweden concluded that hospitalization for RSV bronchiolitis in infancy is highly associated with the development of asthma and allergic sensitization up to age years. 25 It is not clear whether such observed associations are related to predisposition to reactive airway disease or a direct consequence of RSV infection. 7,26 Clearly, this is an area that deserves additional study. A second potential limitation of our study was the finding of significant differences in the distribution of several patient attributes between the study groups, including race, multiple birth, gestational age, and birth weight. In conducting this cost analysis, we made an effort to control for the possible confounding factors to the prophylactic effect of palivizumab as a result of nonrandomization of the Medicaid data by including these patient characteristics in the multivariate regression models. A third potential criticism of our study is the selection of inpatient costs on the basis of RSV diagnoses and test results when the testing procedures are 100% sensitive and the lack of RSV test results for selection of ambulatory care costs. This potential concern was addressed through sensitivity analysis, with outcomes remaining robust with varying cost assumptions. With 4 million live births in the United States each year 27 and 5.6% occurring at 32 to 35 weeks EGA, 28 RSV prophylaxis in this relatively large population subgroup has major economic implications for payers. The absence of significant differences in direct costs of hospitalization and ambulatory care for RSV infection between the study groups without addition of the palivizumab expense presents an interesting dilemma for payers, such as North Carolina Medicaid, in the absence of long-term health outcome data for RSV-infected infants in this gestational age group. In the absence of information on longterm health outcomes and costs and with limited resources available for publicly funded health care, it seems prudent to consider cost of palivizumab as 1 factor in developing policies and practice guidelines for treatment of patients who are at risk for serious RSV infections. Focusing effort on preventive measures to reduce risk of severe RSV infection in higher ARTICLES 1617

8 risk groups, if these can be defined, seems appropriate. Currently, the 2 risk factor approach to identifying at-risk infants seems to be a conservative approach to this issue but was not supported by analyses of our data. Public health efforts to reduce prematurity and low birth weight should be the highest priority. 29 Development of an effective vaccine is also viewed as a high priority. 2,8,20 Promotion of continued breastfeeding throughout infancy has been suggested. 29 Health care providers have a role in counseling caregivers on the importance of not exposing infants to tobacco smoke (including prenatal), good handwashing, and limiting infant exposure to crowds and children with respiratory infections. 7,12 These public health initiatives will likely provide a means to reduce infection but have not been widely implemented. In the past few years, many states have suffered with deficit budgets and ballooning Medicaid costs. 30 As a result, many states have cut enrollment, eliminated services, and decreased reimbursement. 30,31 Drug costs have continued to rise as a percentage of the total expenditures and in total dollars. Palivizumab costs were the highest for any single drug within the AccessCare/Community Care Program in With costs of RSV prophylaxis in our 32- to 35-week EGA study group far exceeding the costs of treatment for RSV-attributed infections, and in the absence of clear data to support long-term outcome differences in this gestational age group with or without palivizumab treatment, consideration must be given to the continued use of this drug in 32- to 35-week EGA infants. When Medicaid programs are faced with decisions concerning cutting enrollment versus reducing drug use that is not costeffective, we feel certain that the choice is obvious. CONCLUSIONS This direct cost analysis of palivizumab prophylaxis in infants who were born at 32 to 35 weeks EGA found that the seasonal costs of RSV prophylaxis in this study group far outweigh the costs of care for RSV-attributed bronchiolitis. In the Synagis prophylaxis group, the per-participant average seasonal cost of RSV prophylaxis represented 97.6% of the RSV-related costs for this group. The computed net cost of prophylaxis needed to prevent 1 RSV hospitalization in this study sample was $ The aggregate cost of palivizumab injections ($ ) in the Synagis prophylaxis group exceeded the costs of all other health care services ($ ) during the study period. Of particular importance was the observation that there were no deaths reported in our cohort during the study period, the same as observed in the IMpact RSV placebo group. 6 Death was not a factor in our cost analysis. ACKNOWLEDGMENTS This research was conducted as a collaboration between AccessCare and the Agency for Healthcare Research and Quality s University of North Carolina Center for Education on Research on Therapeutics (Harry Guess, Center for Education and Research on Therapeutics Principal Investigator; award no. 2 U18 HS ); contractual funding was provided by AccessCare, a not-for-profit organization. We are indebted to the study participants; practice site staff and physicians; and patient support coordinators, local care managers, regional project managers, and data support staff at AccessCare for dedication and involvement in data collection. We are also grateful to Dr Charles Willson, East Carolina University, and Dr Marian Earls, at Guilford Child Health, for support and assistance in study planning and implementation. REFERENCES 1. Centers for Disease Control and Prevention. Bronchiolitis-associated outpatient visits and hospitalizations among American Indian and Alaskan Native children United States, MMWR Morb Mortal Wkly Rep. 2003;52: Centers for Disease Control and Prevention, National Center for Infectious Diseases. Respiratory Syncytial Virus; Available at: Accessed June 10, Boyce TG, Mellen BG, Mitchel EF, Wright PF, Griffin MR. Rates of hospitalization for respiratory syncytial virus infection among children in Medicaid. J Pediatr. 2000;137: Sorrentino M, Powers T, and the Palivizumab Outcomes Study Group. Effectiveness of palivizumab: evaluation of outcomes from the 1998 to 1999 respiratory syncytial virus season. Pediatr Infect Dis J. 2000;19: Shay DK, Holman RC, Newman RD, Liu LL, Stout JW, Anderson LJ. Bronchiolitis-associated hospitalizations among US children, JAMA. 1999;282: The Impact-RSV Study Group. Palivizumab, a humanized respiratory syncytial virus monoclonal antibody, reduces hospitalization from respiratory syncytial virus infection in high-risk infants. Pediatrics. 1998; 102: Pickering LK, ed. Red Book: 2003 Report of the Committee on Infectious Diseases. 26th ed. Elk Grove Village, IL: American Academy of Pediatrics; Kamal-Bahl S, Doshi J, Campbell J. Economic analyses of respiratory syncytial virus immunoprophylaxis in high-risk infants. Arch Pediatr Adolesc Med. 2002;156: Howard TS, Hoffman LH, Stang PE, Simoes EA. Respiratory syncytial virus pneumonia in the hospital setting: length of stay, charges, and mortality. J Pediatr. 2000;137: Shireman TI, Braman KS. Impact and cost-effectiveness of respiratory syncytial virus prophylaxis for Kansas Medicaid s high-risk children. Arch Pediatr Adolesc Med. 2002;156: American Academy of Pediatrics, Committee on Infectious Diseases and Committee on Fetus and Newborn. Policy statement: prevention of respiratory syncytial virus infections: indications for the use of palivizumab and update on the use of RSV-IGIV (RE9839). Pediatrics. 1998; 102: American Academy of Pediatrics, Committee on Infectious Diseases and Committee on Fetus and Newborn. Revised indications for the use of palivizumab and respiratory syncytial virus immune globulin intravenous for the prevention of respiratory syncytial virus infections. Pediatrics. 2003;112: Joffe S, Ray T, Escobar GJ, Black SB, Lieu TA. Cost-effectiveness of respiratory syncytial virus prophylaxis among preterm infants. Pediatrics. 1999;104: Lofland JH, Touch SM, O Connor JP, et al. Palivizumab for respiratory syncytial virus prophylaxis in high-risk infants: a cost-effectiveness analysis. Clin Ther. 2000;22: Farina D, Rodriguez S, Bauer G, et al. Respiratory syncytial virus prophylaxis: cost-effective analysis in Argentina. Pediatr Infect Dis J. 2002;21: Marchetti A, Lau H, Magar R, Wang L, Devercelli G. Impact of palivizumab on expected costs of respiratory syncytial virus infection in preterm infants: potential for savings. Clin Ther. 1999;21: Hennekens CH, Buring JE. Epidemiology in Medicine. Boston, MA: Little, Brown and Company, Anderson RM, McCutcheon A, Aday LA, Chiu GY, Bell R. Exploring dimensions of access to medical care. Health Serv Res. 1983;18: Aday LA, Anderson R. A framework for the study of access to medical care. Health Serv Res. 1974;9: Viswanathan M, King V, Bordley C, et al. Management of bronchiolitis in infants and children. Evidence Report/Technology Assessment No COST ANALYSIS OF PALIVIZUMAB PROPHYLAXIS

9 69 (Prepared by RTI International-University of North Carolina at Chapel Hill Evidence-Based Practice Center under Contract No ). Rockville, MD: US Department of Health and Human Services, Agency for Healthcare Research and Quality; 2003 (AHRQ Publication No. 03-E014) 21. Kraemer HC, Thiemann S. How Many Subjects? Statistical Power Analysis in Research. Newbury Park, CA: Sage Publications; Schlesselman JJ. Sample size requirements in cohort and case-control studies of disease. Am J Epidemiol. 1974;99: Gujarati DN. Basic Econometrics. 2nd ed. New York, NY: McGraw-Hill Book Company; Kennedy P. A Guide to Econometrics. 2nd ed. Cambridge, MA: MIT Press; Sigurs N, Bjarnason R, Sigurbergsson F, Kjellman B. Respiratory syncytial virus bronchiolitis in infancy is an important risk factor for asthma and allergy at age 7. Am J Respir Crit Care Med. 2000;161: Fonseca CB, Grisi S. Bronchiolitis, respiratory syncytial virus, and recurrent wheezing: what is the relationship? Rev Hosp Clin Fac Med Sao Paulo. 2003;58: Division of Public Health, Department of Health and Human Services, State Center for Health Statistics, North Carolina. North Carolina Health Statistics Pocket Guide Raleigh, NC: Division of Public Health, Department of Health and Human Services, State Center for Health Statistics; Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F, Munson ML. Division of Vital Statistics. Births: final data for Nat Vital Stat Rep. 2003; Holman RC, Shay DK, Curns AT, Lingappa JR, Anderson LJ. Risk factors for bronchiolitis-associated deaths among infants in the United States. Pediatr Infect Dis J. 2003;22: Boyd DJ. The bursting state fiscal bubble and state Medicaid budgets. Health Aff. 2003;22: Boccuti C, Moon M. Comparing Medicare and private insurers: growth rates in spending over three decades. Health Aff. 2003;22: WHY WE HAVE A FLU VACCINE SHORTAGE Two decades ago, when Washington embraced the for-profit model to curb escalating charges, health care spending represented 10.5% of gross domestic product. Now it is approaching 16%. We spend more per capita on health care than any other developed country. Yet on the important yardsticks, like life expectancy measured in healthy years, we don t even rank among the top 20 nations. In fact, according to the World Health Organization, we come in an embarrassing 29th, sandwiched between Slovenia and Portugal....The expansion of the abysmal record is one that politicians decline to discuss. The market functions wonderfully when we want to sell more cereals, cosmetics, cars, computers, or any other consumer product. Unfortunately, it doesn t work in health care, where the goal should hardly be selling more heart bypass operations. Instead, the goal should be to prevent disease and illness. But the money is in the treatment, not prevention, so the market and good health care are at odds. Just how much at odds is seen in the current shortage of flu vaccine, as men and women in their 80s and 90s line up for hours at a time, hoping to get the shot they need, but may not receive because not nearly enough has been manufactured....the reason for the shortage is this: preventing a flu epidemic that could kill thousands is not nearly as profitable as making pills for something like erectile dysfunction, a decidedly nonfatal condition. Viagra, for example, brings in more than $1 billion a year for its maker, Pfizer. The profits to be made from selling flu vaccine are measly in comparison. Barlett DL, Steele JB. New York Times. October 24, 2004 Noted by JFL, MD ARTICLES 1619

10 Direct Cost Analyses of Palivizumab Treatment in a Cohort of At-Risk Children: Evidence From the North Carolina Medicaid Program Steven Wegner, Julie Jacobson Vann, Gordon Liu, Patricia Byrns, Clement Cypra, William Campbell and Alan Stiles Pediatrics 2004;114;1612 DOI: /peds Updated Information & Services References Citations Subspecialty Collections Permissions & Licensing Reprints including high resolution figures, can be found at: tml This article cites 22 articles, 9 of which can be accessed free at: tml#ref-list-1 This article has been cited by 12 HighWire-hosted articles: tml#related-urls This article, along with others on similar topics, appears in the following collection(s): Infectious Disease & Immunity disease Information about reproducing this article in parts (figures, tables) or in its entirety can be found online at: ml Information about ordering reprints can be found online: PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly publication, it has been published continuously since PEDIATRICS is owned, published, and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois, Copyright 2004 by the American Academy of Pediatrics. All rights reserved. Print ISSN: Online ISSN: